Key points at a glance
- Specify outcome and predictors before fitting.
- Inspect data quality and model diagnostics.
- Report coefficients with units and uncertainty, without overclaiming cause.
Turn the question into a model
For example, ask whether study time relates to exam performance after accounting for semester. Identify the outcome and predictors from theory or design. Adding every available column does not make an analysis stronger.
Inspect the data and fit
Review missing values, unusual cases, the shape of relationships and residuals. With several predictors, inspect strong overlap among them. Use diagnostics appropriate to your model; a single global test cannot replace informed graphical checks.
Translate coefficients into the subject matter
A coefficient estimates outcome change per unit of a predictor, conditional on the other model variables. State units, direction, intervals and an appropriate measure of model fit. A precise estimate may still have little practical importance.
Limit interpretation to the design
Unmeasured factors can explain an association in observational data. Write “is associated with” unless your design supports a causal inference. Be cautious when generalising outside the sampled population.
Final submission context
Formal details can feel like a separate writing task, but they matter just as much for a printed submission. Anything missing, misplaced or inconsistently formatted in the document will also appear in the bound copy.
Use this guide together with your cover page, table of contents, page numbers, source notes and appendices. Prepare the final PDF for printing and binding only after the complete file has been checked.
Practical check before PDF export
- Do headings, chapter structure and page numbers match?
- Are sources, figures and tables included completely?
- Are required elements such as declarations, appendices or the cover page included where required?
- Did you open and check the final PDF after exporting it?
Checklist
- Outcome and predictors justified.
- Missingness and residuals examined.
- Coefficients reported with units and intervals.
- Causal language matched to design.
Common mistakes
- Selecting variables only by p-values.
- Claiming assumptions were met without diagnostics.
- Presenting association as proof of cause.
Prepare the finished thesis for submission
Once the content and university requirements are checked, review your PDF and the available printing and binding options.
Review printing optionsRelated content
Explore related guidance on thesis structure, sources, formatting and final submission.
Frequently asked questions
Must every variable be normally distributed?
No. Assumptions concern the model and errors, not blanket normality of each raw variable.
What does a coefficient mean?
An estimated outcome change per predictor unit under the model conditions.
Does regression prove causation?
No, not without an appropriate design and assumptions.